Who Actually Owns AI Transformation? Not Just the C-Suite

Companies are spending millions on centralized AI structures and still cannot show meaningful results, argues Tanya Moore, chief people officer of the consulting firm West Monroe. Her diagnosis: the ownership question is being answered wrong.

Ask almost any professional who owns their organization's AI transformation, and they will point to a role at the top of the org chart or a team charged with the work. Moore says those answers are not wrong, but they are incomplete. "Leaders can set direction, but no executive can see the thousands of small workflows that make up company operations," she writes in a recent essay. "The people who understand a process well enough to redesign it are usually the ones executing it every day."

The distinction Moore draws is between users and builders. A user leverages AI to summarize a document or draft an email. A builder looks at a workflow, asks, "Why are we still doing it this way?" — and then changes it.

Employee-led innovation

Builders are not always engineers, and they likely do not report to the chief AI officer. Moore describes builder as a mindset, not a role. Builders share three characteristics: they are close to the work, naturally curious, and they proactively improve workflows without waiting for permission.

The archetype cuts across functions. A builder might be a salesperson redesigning follow-up to increase win rates, an accountant building automation to flag anomalies before month-end close, or an analyst eliminating transactional work to spend more time delivering insights to the business.

Leadership's role, in Moore's framing, is to chart the AI strategy and provide training, tools, guardrails, and agency. Builders distributed across teams and levels, she argues, can scale innovation faster and more effectively than any top-down mandate.

She sees a precedent. "We didn't teach a small group of employees to use the internet," Moore writes of the 1990s-era rollout. "We provided access, training, and guardrails. Eventually, it's how work got done."

Lead with guardrails

Moore warns that leaders can inadvertently get in the way. In an effort to manage AI's very real risks, companies can create so many approval processes that employees stop experimenting altogether. Her prescription: set the non-negotiables around security, privacy, approved tools, and human oversight — then give people room to build.

Permission alone is not enough. People need to see what good looks like, and here peers can be more powerful than a leadership mandate. "A colleague showing how they eliminated three hours of work from their week will do more to make AI approachable than another leadership presentation about its potential," Moore writes.

The leader's job description itself is shifting, she argues. Historically, leaders identified and decided what needed to change. With AI, the job is to create an environment where employees who know the work best redesign work themselves — set direction, establish boundaries, remove barriers, and recognize what works so it can scale.

Good for business and employees

Moore also presses a two-way obligation. If companies ask employees to use AI to make the business more productive, they should also invest in making those skills beneficial to employees' careers.

Companies should expect improved productivity, better financials, and better client outcomes from AI investments, she writes. But employees should come out of the transformation with new skills that help them keep thriving as roles evolve. "Employees need to feel and see the personal and professional benefits AI can provide their careers," Moore writes.

Track metrics that matter

The measurement gap compounds the ownership problem. As organizations implement AI, many are measuring activity instead of outcomes, Moore argues.

Log-ins, licenses, and usage can tell you whether people are accessing the tools. They cannot tell you the impact. Moore offers concrete examples of what counts: a redesigned workflow that gets a product out the door five days faster; a project that requires 40% fewer hours and improves margin; teams creating capacity that can be redirected to clients or higher-value work.

The most meaningful metrics, she writes, are already on the company or board scorecard: revenue growth, margin expansion, client NPS, staffing efficiency, and productivity.

So who owns AI transformation?

Moore's final answer splits the mandate. Leaders own the strategy. Technology teams own platforms and guardrails. But no one person or team can redesign every workflow across an organization. That is where builders come in — the people thinking about how to leverage AI to make work better.

For executives budgeting millions into centralized AI programs, the implication is direct: the return on that spend depends less on the org chart and more on whether rank-and-file employees get the tools, boundaries, and incentives to rebuild their own workflows.